基于遗传神经网络的冷连轧机轧制压力模型  

Rolling Force Models of Cold Tandem Rolling Mill Based on Genetic Neural Networks

作　　者：张大志[1] 程秉祥[2] 李谋渭[3] 孙一康[2] 管克智[3] 

机构地区：[1]北京科技大学高效轧制国家工程研究中心,北京100083 [2]北京科技大学信息工程学院,北京100083 [3]北京科技大学机械工程学院,北京100083

出　　处：《北京科技大学学报》2000年第4期384-388,共5页Journal of University of Science and Technology Beijing

摘　　要：对鞍钢冷轧厂四机架冷连轧机轧制压力模型进行了认真分析，指出了其存在的缺陷．把遗传算法（GeneticAlgorithms，简称GA）和神经网络有机结合，设计出了具有遗传算法性能参数优选、网络结构参数优选、网络性能参数优选以及GA-BP算法联合进行网络权值修改几种功能的遗传神经网络，建立了基于遗传神经网络的新冲连轧机轧制压力模型．通过原模型计算值、新模型计算值与实测值之间的对比分析可知，遗传神经网络模型计算精度优于传统轧制力模型．Some defects of the traditional rolling force models of cold tandem rolling mill were found out, and new rolling force models based on genetic neural networks were set up. The comparison results of the measured rolling force of cold tandem rolling mill with the calculated value from the traditional models and also with the calculated value from the new rolling force models based on genetic neural netWothe show thatthe calculating precision of the new models is better than that of the traditional models.

关 键 词：BP算法 冷连轧机 轧制压力模型 遗传神经网络 

分 类 号：TG333[金属学及工艺—金属压力加工] TG335.12